Over the past 50 years, health and functional status of the American population has continued to improve but these improvements have not been equally distributed across races, ethnicities and socioeconomic groups. Perhaps nowhere is the challenge to reduce health disparities more apparent than in Alzheimer's disease (AD) and other dementias. Risk of AD is higher among ethnic minorities, and because many public and private insurers do not cover institutionalized or paid home health care services that individuals with these diseases require, the burdens faced by family caregivers from diverse racial, ethnic and socioeconomic backgrounds is high. The goals of this project are two-fold. First, we will estimate predictive models of cognitive decline, AD onset, other health and functional status changes of middle age and older Americans. We model caregiving for person with AD and analyze and quantify how health and other indicators of well being evolve over the life course for caregivers of family members with AD and contribute to health disparities. Second, we will utilize a dynamic micro simulation model of health, cognitive decline and economic outcomes to quantify health of future populations and burden of AD across diverse groups. Our ability to achieve the proposed aims is enhanced by use of the best data sets available for this research and that are established NIH-funded longitudinal cohorts: the Health and Retirement Study and its sub-study the Aging Demographics and Memory Study. It is further enhanced by a team of researchers with expertise in both dementia and microsimulation.